Medicare utilization and spending vary substantially across regions of the US (Fisher 2003a, 2003b), and the additional spending in high-cost areas has little known value (Fisher et al., 2009). Far less is known about the causes of such variations. Two related, prevailing hypotheses are that local norms create a region's signature style, and those high spending results from too many hospitals and physicians. Under these hypotheses, Medicare and private spending are predicted to have a high degree of similarity within a given region. Under these hypotheses, the higher levels of spending represent pure waste in Medicare and in the US health care system as a whole. A number of policies are justified by these hypotheses, including restricting the number of doctors and hospitals and reducing Medicare reimbursement levels in high spending areas. In this study we propose and test a different theoretical explanation of Medicare's regional variations. We accomplish this by using a rich, newly-available dataset in conjunction with other data to test a widely- familiar economic theory of provider behavior. Specifically, we will use data from FAIR Health, Thomson Reuters MarketScan, and Medicare Part B to test whether geographic variations in Medicare spending for physicians can be explained by the profitability of Medicare relative to other payers. We develop this hypothesis based on the economic model of stepwise demand. The central insight from this model relevant to Medicare variations is straightforward: Medicare's profitability relative to other payers affects Medicare utilization and spending. This occurs because providers respond to Medicare's profitability by altering their payer mix. Providers accomplish this by altering the prices they charge to privately-insured patients. Contrary to cost shifting, this yields the hypothesis that reductions in Medicare fee changes would be met with reductions in prices for privately-insured patients. Also contrary to cost offsets, this hypothesizs that reductions in Medicare reimbursement would lead to lower, not higher, Medicare utilization. This project will achieve three specific aims. In Specific Aim 1, we will test for cost shifting. Or preliminary studies contradict the cost-shifting hypothesis and support the hypothesis from the stepwise demand model. This heightens the need to complete Specific Aims 2 and 3. In Specific Aim 2 we will test for cost offsets. In Specific Aim 3 we examine whether and to what extent the regional variations in Medicare spending and utilization are due to physicians altering their payer mix in response to Medicare's profitability relative to privately-insured patients. This will establish what percent of the regional variation in Medicare spending is due to differences in the profitability of Medicare relative to privately-insured patients. The broad, long term objective o this study is to guide researchers and policy makers' understanding of physicians' responses to changes in Medicare reimbursement and the source of Medicare variations. These results will inform which policies, if any, should be implemented to reduce these variations.